Use of EnzymeArtiﬁcial Cells for Genetic Enzyme … Enzyme Defects that Increase Systemic Substrates toToxic Levels ... enzyme may result in immunological and ... enzyme catalase

CHAPTER 6

Use of Enzyme Artificial Cells forGenetic Enzyme Defects thatIncrease Systemic Substratesto Toxic Levels

6.1. Introduction

General

The basic principle on the use of artificial cells containing enzymesin enzyme therapy has been discussed in Chapter 2 (Fig. 2.3). Thisprinciple has been studied for use in enzyme therapy for geneticenzyme defects, removal of unwanted metabolites and treatmentof amino acid-dependent tumors (Fig. 6.1). This chapter describesexamples of the use of enzyme artificial cells in three genetic enzymedefects: Acatalasemia, Lesch-Nyhan disease and phenylketonuria(Fig. 6.1). Other terminologies for genetic enzyme defects includeinborn errors of metabolism, hereditary enzyme defects and congenitalenzyme deficiency. The cause of each of these conditions is thecongenital defects of an enzyme or an enzyme system. This results inthe accumulation of the substrate for the enzyme to an unacceptableor toxic level.

Substrate→ // DEFECTIVE ENZYME //→ Product

The accumulation can be inside the cells as in the case of storagediseases, or throughout the body fluid. There is already an extensiveliterature on inborn errors of metabolism (Scriver, 1989); thus, this

130

Use of Enzyme Artificial Cells for Genetic Enzyme Defects 131

ENZYME ARTIFICIAL CELLS

ENZYMESSUBSTRATES PRODUCT

ANTIBODYWBCTRYPTIC ENZYMES

IN CONGENITAL ENZYME DEFECTS

ACATALASEMIA

CATALASEPEROXIDE

XANTHINEOXIDASE

HYPOXANTHINE

LESCH-NYHAN DISEASE PHENYLKETONURIA

PhenylalanineAmmonia

Lyase

PHENYLALANINE

ASPARAGINASEASPARAGINE TYROSINASETYROSINE

AMINO ACID DEPENDENT TUMOURS

GENETIC ENZYME DEFECTS

CELL

ARTIFICIAL CELL

IMPLANTED: to work with body cells INJECTION

BODY FLUID(Blood, plasma etc)

PATIENT

Retained in chamber outside body

GivenOrally

ExcreteIn stool

Given orally to travel through intestineas microscopic dialysers & bioreactors

4 ROUTES OF ADMINISTRATION STUDIED

Fig. 6.1. Upper : Examples of enzyme artificial cells. Lower : Routes ofadministration.

132 Artificial Cells

chapter will concentrate on the type where the substrates accumulatethroughout the body fluids.

Enzyme replacement therapy

Much study has been carried out on the use of enzyme therapyin treating defective enzymes. However, as discussed in Chapter 2,the use of enzymes directly has a number of problems: a) the enzymemust be able to be located at the site of action; b) the enzyme has to beavailable in a highly purified and non-toxic form; c) the heterogeneousenzyme may result in immunological and hypersensitivity reactions;and d) the enzyme has to act for a sufficient length of time. Some ofthese problems might be avoided by using artificial cells containingthe defective enzyme (Chang, 1964, 1965; Chang et al., 1966;Chang & Poznznsky, 1968) (Fig. 6.1). In this way, the enclosedenzyme does not leak out and become involved in immunologicalor hypersensitivity reactions, but can act on the permeant substrates.We investigated the use of enzyme artificial cells for three types ofcongenital enzyme defects: Acatalasemia, Lesch-Nyhan Disease andphenylketonuria using respectively, artificial cells containing catalase,xanthine oxidase and phenylalanine ammonia lyase (Fig. 6.1). Fourroutes of administrationwere studied (Fig. 6.1) i.e. the enzyme artificialcells are: 1) implanted into the body to work alongside the bodycells; 2) retained in chambers outside the body to act on the bodyfluid perfusing the chamber; 3) injected; or 4) given orally so that asthe enzyme artificial cells move through the lumen of the intestine,they will act as microscopic bioreactors on the substrates entering theintestinal lumen from the body. In this way, the enzyme artificial cellswould not be retained in the body as they would be excreted in thestool after completion of their function.

6.2. Acatalasemia: Congenital Defect inthe Enzyme Catalase

The first study is to use the simplest animal model to determine ifenzyme therapy using artificial cells is feasible, effective and free


from immunological effects. Acatalasemia is a congenital defect in theenzyme catalase leading to a deficiency in the removal of its substrateperoxide. This is a rare condition in humans, but there is a convenientstrain of mice with congenital defect in the enzyme catalase (Feinsteinet al., 1966).These mice have a blood catalase activity of only 1–2% ofnormal mice and a total body catalase activity of 20% of normal mice.Furthermore, the enzyme catalase can be obtained easily in a highlypurified form.Our initial study shows that injection of catalase artificialcells into acatalasemic mice can effectively remove peroxides (Chang& Poznansky, Nature, 1968b). We therefore carried out further studiesto determine the in vivo kinetics, effectiveness and immunological andhypersensitivity effects (Chang, 1971b, 1972f; Poznansky & Chang,1974). The details are as given in the next section.

6.3. Preliminary Study on the Use of CatalaseArtificial Cells to Replace DefectiveCatalase in Acatalasemia

The first study is to answer the question “Can implanted enzymeartificial cells work alongside the body cells to replace their enzymedefects?” (Fig. 6.1).

Implantation of catalase artificial cells

Collodion membrane catalase artificial cells were prepared asdescribed in the Appendix under “Methods of Preparation of ArtificialCells.” In acatalasemic mice which are injected with sodium perboratewithout the protection of a previously injected catalase, the red colorof their pupils turned brown and the animals became increasinglyimmobile, so that by 20min they became flaccid and were inrespiratory distress. Acatalasemic mice protected by an intraperitonealinjection of catalase artificial cells were slightly immobile 5min afterthe injection of perborate, and the color of their pupils darkened.However, in 10min, the mice became more active and their pupilsreturned to their normal red color, and except for one animal, they


were moving about freely at 20min. Acatalasemic mice injected witha catalase solution and the normal control mice recovered even faster.Figure 6.2 shows that within 20min after injection of perborate,

normal mice could remove most of the perborate, leaving only2.5% ± 2.5% S.D. in the body. Acatalasemic mice receiving nocatalase injection could only remove a small amount of the injectedperborate, leaving 70% ± 8.2% S.D. after 20min. Acatalasemic micereceiving either catalase artificial cells or catalase solution, couldmore efficiently remove the injected perborate, leaving respectively,16% ± 3.5% S.D. and 7% ± 3.6% S.D. 20min after injection. Thus,intraperitoneally injected liquid catalase and catalase artificial cellsboth acted efficiently in replacing the enzyme defect. Catalase insolution is expected to be more efficient since there is no membranebarrier for the diffusion of the substrate. On the other hand, catalase insolution may give rise to immunological and hypersensitivity reactionsand these will be studied in a later section. In the case of catalase

Control

Control

Catalase

Catalaseartificial cells

Control

Catalase artificial cells

I.P implantation Shunt

Normal mice Acatalasemic mice

%

Pe

rbo

rate

re

co

ve

red

100

0% Total body perborate recovered 20 min after

injection of perborate. (Chang & Poznansky, Nature, 1968.)

Fig. 6.2. Implanted catalase artificial cells remove peroxide from the body.They are also effective in removing peroxide from the body when retainedin a chamber outside the body and perfused with peritoneal fluid. (Modifiedfrom Chang & Poznansky, Nature, 1968b.)


artificial cells, there was no increase in blood catalase level 20minafter injection, showing that there is no leakage of catalase into thesystemic circulation.

Catalase artificial cells retained in a chamber forbody fluid perfusion

In the second set of experiment, to avoid the introduction andaccumulation of catalase artificial cells, especially in long-term use,we carried out the following study.We retained catalase artificial cellsin a chamber outside the body and perfused these cells by recirculatingperitoneal fluid from the animal (Fig. 6.1). Here the animals wereanesthetized with Nembutal. Each animal received an intraperitonealinjection of 4.5ml of inperinol fluid (Abbott Laboratory). This fluid wascontinuously recirculated through a smaller modified version of theextracorporeal shunt chamber (Chang, 1966) for 20min after injectionof perborate. In one group of animals, the shunt contained 4ml of a50% suspension of catalase artificial cells. In the control group, nocatalase artificial cells were present in the shunt. Figure 6.2 shows thatin the control group, the perborate remaining in the body was 76.1%±13.4% S.D. after 20min. In the chamber containing catalase artificialcells, only 32% ± 9.8% S.D. of the injected perborate was recovered20min after injection.

6.4. In Vivo Kinetics of Catalase Artificial Cells forAcatalasemic Mice

Summary

The enzyme kinetics and immunological properties of collodionartificial cells containing catalase (H202: H202 oxidoreductase, EC1.11.1.6) were studied and compared with catalase in free solution.In vitro studies showed that the Km for both forms of catalase was0.55M. The Vm of the catalase artificial cells was 0.5mM/min, whilethat of catalase in free solution was 2.5mM/min. In vivo studies onacatalasemicmice showed that catalase artificial cells were as effectiveas the same assayed amount of catalase in free solution in reducing total


body substrates. After injection, the catalase activity of the artificialcells in the body had a half-time of 4.4 days. This is more stable thancatalase in free solutionwhich had a half-life of 2.0 days after injection.In vitro studies showed that antibody to catalase would not enter theartificial cells. Immunizing doses of catalase artificial cells did notproduce antibodies. Immunizing doses of catalase solution resulted inthe production of antibodies to catalase. Injection of catalase solutioninto these immunized mice resulted in anaphylactic reactions. On theother hand, catalase artificial cells injected into these immunizedmicecontinued to act without causing any adverse reactions.

Methods of preparation of catalase artificial cellsand experimental procedures

These are described in detail in Appendix II.

Characteristics of acatalasemic mice

Experiments showed that the acatalasemic mice bred in this laboratoryfrom the original Feinstein strain were the same as the original strain.Thus, the blood and total body catalase levels were respectively, 1–2%and 15%, of the normal mice.

Kinetics of recovery of injected perborate

Figure 6.3 represents the time course of removal of injected perboratefrom acatalasemic mice and from normal mice. Normal mice removedinjected perborate rapidly, whereas acatalasemic mice could onlyremove it at a very slow rate. While the normal mice showed nooutward symptoms after the injection of perborate, the acatalasemicmice were in some distress. The eyes of these acatalasemic micechanged from red to dark brown due to methemoglobin formation,and they became increasingly immobile and by 20min most of themwere flaccid and in respiratory distress. Quantitative analysis showedthat injection of either catalase solution or catalase artificial cells couldsignificantly increase the rate of removal of the injected perborate(Fig. 6.3). Acatalasemic mice given prior intraperitoneal injections of


Fig. 6.3. Rate of removal of peroxide in nonimmunized rats using PALartificial cells as compared to catalase solution and control saline. (Modifiedfrom Poznansky & Chang, 1974.)

either catalase in free solution or catalase artificial cells did not sufferany distressful symptoms and were able to remove the peroxide ina manner similar to the normal mice. Thus, both catalase solutionand catalase artificial cells were effective and safe after the firstimplantation.

6.5. Immunological Studies on Catalase Artificial Cellsin Acatalasemic Mice

Effects of immunizing doses of catalase solution andcatalase artificial cells on antibody production

The next step is to study whether there is a difference in theimmunological response when catalase solution and catalase artificialcells, respectively were administered. Acatalasemic mice and rabbits


Fig. 6.4. Ouchterlony double-diffusion technique showed that antibodiesto catalase were produced in the acatalasemic mice and rabbits immunizedwith catalase in free solution. When catalase artificial cells were used, thesame test showed no detectable antibodies.

were immunized over a period of 90 days. Each immunizing doseconsisted of Freunds’ adjuvant with catalase solution or catalaseartificial cells (0.1ml containing 4.8 perborate units of catalase).Ouchterlony double-diffusion technique showed that antibody titerswere produced in the acatalasemic mice and rabbits immunized withcatalase in free solution (Fig. 6.4). The same test showed that nodetectable antibody titers were formed in those animals receivingimmunizing doses of catalase artificial cells (Fig. 6.4)

Permeability of membrane of collodion artificial cellsto catalase antibody

Catalase antibodies were incubated with catalase artificial cells orwith control artificial cells containing no catalase. No antibodieswere detected inside the artificial cells when tests were carried out


using either the radioactive labeled antibody technique or the micro-complement fixation technique. This suggests that the membrane ofartificial cells is not permeable to catalase antibodies.

Response of immunized acatalasemic mice to injections ofcatalase artificial cells or catalase solution

Intraperitoneal injection of catalase solution into the immunizedacatalasemic mice caused them to go into anaphylactic shock and80% of them died within the first 10min (Fig. 6.4). Death wasaccompanied by violent gasping, convulsions and collapse similar tothat observed in anaphylactic shock in mice. In the mice that survived,the injected catalase solution was not as effective in removing theperoxide (Fig. 6.5) Injection of catalase artificial cells did not produce

Fig. 6.5. Acatalasemic rats receiving immunizing doses of catalase solutionto produce antibody to catalase. This was then followed by the implantationof either catalase solution or catalase artificial cells. (Modified fromPoznansky & Chang, 1974.)


any adverse reactions. In addition, Fig. 6.5 shows that catalase artificialcells removed peroxide at the same rate as when the cells were injectedinto the nonimmunized acatalasemic mice (Fig. 6.3).

Appearance of catalase activity in the blood afterinjection of catalase solution or catalase artificial cells

Intraperitoneally injected catalase solution is expected tomove quicklyinto the blood stream. However, similarly injected catalase artificialcells should stay inside the peritoneal cavity where they act on theperborate diffusing into the artificial cells. Thus, the enzyme shouldnot leak out of the artificial cells into the circulating blood. To analyzethis, the amount of catalase in free solution or of catalase artificialcells was reduced to one-third. This reduces the mortality rate from80 to 20% for the immunized mice receiving the catalase solution. Inthis way, the kinetics of removal of the perborate could be studied.The result shows that free catalase solution is not as effective ascatalase artificial cells in removing the peroxide in immunizedmice.Toanalyze the reason for this, the following study was carried out in theseanimals.Figure 6.6 shows the appearance of catalase in the bloodstream

of immunized and nonimmunized acatalasemic mice following theintraperitoneal injection of catalase solution or catalase artificial cells.There was no detectable leakage of catalase into the blood streamafter the intraperitoneal injection of catalase artificial cells. There wasno significant change in the blood catalase level when followed forup to 3 months. The artificial cells with its catalase are retained inthe peritoneal cavity after injection. On the other hand, injectionof catalase solution into acatalasemic mice resulted in the rapidappearance of the enzyme in the blood stream. The peak catalaseactivity in the blood reached 182.8 perborate units 8 h after injection ofthe catalase solution into nonimmunized acatalasemic mice (Fig. 6.6).What is surprising is that when the same catalase solution was injectedinto immunized acatalasemic mice, the blood catalase level reacheda peak of only 19.6 perborate units 4 h after the injection. This is mostlikely due to the formation of antigen-antibody complexes that arequickly removed.


Fig. 6.6. Appearance of catalase activity in the circulating blood afterintraperitoneal injection of catalase solution or catalase artificial cells.After intraperitoneal injection, catalase in solution moves rapidly into thecirculation catalase artificial cells remains inside the artificial cells and doesnot leak into the circulation (From Poznansky & Chang, 1974.)

6.6. Conclusion from Results of Basic Study Usingthe Acatalasemic Mice Model

On a theoretical level, this basic study supports the proposal of usingenzyme artificial cells to replace the defective enzyme (Fig. 6.1). Thus,artificial cells containing an enzyme catalase can effectively replacethe defective catalase enzyme in this animal model. The result alsoshows that catalase inside artificial cells is prevented from giving riseto immunological reactions. Furthermore, the membrane of artificialcells is not permeable to catalase antibodies in mice previouslyimmunized with catalase solution. Therefore, the catalase artificial


cells can continue to function in those acatalasemic mice previouslyimmunized against catalase. Therefore, in theory, enzyme artificialcells should allow for enzyme therapy in congenital enzyme defectswhere the accumulation of substrate is in the body fluid.However, at that time, enzymes for the most common genetic

enzyme defect, phenylketonuria, could only be extracted from theliver. These liver enzymes are very complex and unstable and notsuitable for use in artificial cells for enzyme therapy. With thelater availability of simple and stable enzymes from microorganisms,enzyme therapy becomes muchmore feasible and practical. However,the major barrier is that enzymes are not stable at body temperatureand thus can only function in the body after injection for a few days.In congenital enzyme defects, the duration of treatment is for the lifeof the patients. Repeated injection of enzyme artificial cells over themany years of treatment would result in the accumulation of muchforeign material in the body. To solve this problem, catalase artificialcells retained in small chambers outside the body and perfused by bodyfluid can also remove perborate. However, a more convenient routeof administration would be better for pediatric patients.We, therefore,looked at giving enzyme artificial cells orally (Fig. 6.1).This is based onour earlier finding of the safety and effectiveness of oral administrationof urease artificial cells in rats to remove systemic urea (Chang, 1972a).The safety of oral urease artificial cells has later been demonstrated inclinical trials in kidney failure patients (Kjellstrand et al., 1981).The next two studies will therefore, look at the use of oral

administration. This way, the enzyme artificial cells move down theintestinal lumenwhere they act as combinedmicroscopic diaysers andenzyme bioreactors to remove unwanted substrates moving into thelumen from the body fluid. After completion of their functions, they areexcreted in the stool without being accumulated in the body (Fig. 6.1).The enclosed enzymes are protected from intestinal tryptic enzymes.

6.7. Oral Xanthine Oxidase Artificial Cells ina Patient with Lesch-Nyhan Disease

Lesch-Nyhan disease is caused by a genetic defect in the enzymehypoxanthine phosphoribosyltransferase (HPRT) (Wilson et al., 1983).


This leads to overproduction of purine and accumulation of oxypurineintermediates and uric acid. It has been suggested that damage to thebrain in Lesch-Nyhan disease may be secondary to the accumulationof oxypurines such as hypoxanthine (Lloyd et al., 1981). The enzymeHPRT is very complex and difficult to extract on a large scale. However,there is an enzyme, xanthine oxidase, produced by fermentation andcan be easily purchased. We, therefore, look into giving xanthineoxidase (XOD) artificial cells orally (Chang, 1989a; Palmour, Changet al., 1989). As they move through the intestine, the XOD artificialcells act as amicroscopic combined dialyser-enzyme reactor (Fig. 6.1).The enclosed XOD is protected from tryptic enzymes, but theultrathin selectively permeablemembranes allowpassive entry of smallsubstrate molecules like hypoxanthine. Hypoxanthine is highly lipidsoluble, and can therefore equilibrate rapidly between the body fluidand the intestine. In this way, hypoxanthine and other oxypurines fromthe body fluid entering the intestinal lumen could be converted intouric acid. The uric acid in the intestine would be excreted in the fecesdirectly or as allantoin.Thismay result in the depletion of the oxypurinebody pools. To test this idea, we administered XOD artificial cells to apatient with Lesch-Nyhan disease.Collodion membrane artificial cells containing XOD were prepared

using the updated method as described in Appendix II under“Preparation of Enzyme Artificial Cells.” XOD was obtained fromBoehringer, Montreal (grade III from buttermilk, chromatographicallypurified, 50U at 1–2U/mg protein). 1ml artificial cells contained 5UXOD. The XOD artificial cells were stored as a 50% suspension insaline solution at 4◦C for 0–3 days before use. Since the artificial cellsare destroyed by pH of less than 3.5, it is important to administer thecells together with a buffer.

Volume ratio of artificial cells/substrate solution

In order to analyze the amount of XOD artificial cells needed fororal administration, the following in vitro study was carried out. Freshsubstrate solutions were prepared containing 40μM hypoxanthinein phposhate buffer (0.1M), pH 7.5. Four different volume ratios ofartificials/substrate solution were studied, i.e. 1/10, 1/100, 1/500 and


1/1000. Hypoxanthine was converted by xanthine oxidase into uricacid and the reaction was followed by the rate of formation of theproduct, uric acid.

Effects of pH on enzyme activity

For oral administration, it is important to analyze the relationshipbetween pH and enzyme activity. It is also important to analyze thestability of the collodion artificial cells and their contents at differentpH environments. The pH’s of the substrate solutions were adjustedusing citric-phposphoric acid-borate-HCl buffer. The following pH’swere studied: pH 2.5, 3.5, 4.5, 5.5, 6.5, 8.5, 9.5 and 10.5, there wasleakage of all the contents of the artificial cells due to destructionof the membrane. At pH 3.5 and above, there was no leakage. Thisaspect is important in the clinical use of the artificial cells by oraladministration. The artificial cells have to be buffered in order toprevent their destruction in the very acidic stomach environment.The enzyme activity of artificial cells containing xanthine oxidase isoptimal in the pH range of 8.5 to 9.8 that is normally present in theintestinal lumen.

Patient with Lesch-Nyhan disease

The patient, a 29-month-old boy with severe Lesch-Nyhan disease(HPRT activity less than 3 pmol inositic acid/mg protein/h) waswithdrawn from allopurinol therapy over 5 days. Blood and urine wereobtained at 0830 h and 1630h. After baseline studies for 3 days, thepatient was treated with 100 U of XOD artificial cells via a nasogastrictube twice a day (0900 and 1200h) for 10 days. A sample of hiscerebrospinal fluid was obtained at 1630 h before treatment and onthe final day of therapy.Purinesweremeasured by high-performance liquid chromatography

and uric acid and creatinine were measured enzymatically andcolorimetrically, respectively.After withdrawal of the allopurinol, uric acid excretion rose rapidly;

baseline uric acid/creatinine ratios at 0900 h ranged between 3.5


XA

N

INOH

X

UA

XA

N

INO

HX

UA

Ce

reb

ral

Sp

ina

l F

luid

Pu

rin

es

(µ

g/m

l)

0

6

Before After 10 days of oralenzyme artificial cells

UA: uric acidHX: hypoxanthineINO: inosineXAN: Xanthine

Fig. 6.7. Effect of oral XOD artificial cells on cerebral spinal fluidconcentrations of uric acid, hypoxanthine, inosine and xanthine (modifiedfrom Palmour et al., 1989).

and 4.1. With the introduction of XOD artificial cells, the morningand afternoon ratios fell to normal within 3 days. During a febrileepisode of otitis media the ratios rose transiently. Urinary hypoxanthineexcretion also responded promptly to the XOD treatment, the patternof transient escape during the febrile illness being repeated. The CSFhypoxanthine levels fell by 25% after 10 days of XOD artificial celladministration (Fig. 6.7) and the CSF inosine levels fell by 32% (380to 260 ng/ml). The levels of CSF xanthine and uric acid, both normal,remained unchanged.The effect of XODon plasma inosinewas rapid (Fig. 6.8) andwithin 3

days, the plasma inosine levels had fallen to below 5μg/ml. There wasa reduction in plasma xanthine levels after day 6. The hypoxanthinelevels manifested wide diurnal fluctuations but a regression lineindicated a gradual decrease.


Pla

sm

a p

uri

ne

s (

µg

/ml)

0

80

Oral xanthine oxidase artificial cells (days)start

10

xanthine

Inosine

1Pre Px

Fig. 6.8. Plasma xanthine and inosine levels before and after administrationof oral XOD artificial cells on day 1 shown as “start” in the figure.Hypoxanthine levels, not shown here, manifest wide diurnal fluctuations buta regression line indicates a gradual decrease (modified from Palmour et al.;Chang, 1989).

6.8. Conclusion Based on Study

The purpose of our short trial of ten days is only to determine ifadministration of oral XOD artificial cells can change the systemiclevels of the different substrates. As such, the very preliminary resultsuggests the feasibility of pursuing the study further. Much longerclinical trials would be needed to observe the neurological changesor behavioral changes. This is an extremely rare inborn error ofmetabolism, and thus we could not find any commercial interest inmanufacturing this product for use in long term clinical trials. However,the very preliminary demonstration of the safety and effectiveness oforal enzyme artificial cells in a patient has led us to develop this


approach at treating one of the most common genetic enzyme defects,phenylketonuria.

6.9. Phenylketonuria: Genetic Defect in EnzymePhenylalanine Hydroxylase

Phenylketonuria

Phenylketonuria (PKU) is caused by a genetic defect in the liverenzyme, phenylalanine hydroxylase.The resulting increase in the levelof systemic phenylalanine in the first few years of life can lead to severemental retardation (Nylan, 1974; Scriver, 1989). A low phenylalaninediet is, at present, the only effective treatment for phenylketonuria(Scriver, 1989). This diet must begin very early after birth. However,this diet is difficult to follow and does not prevent elevation of bloodphenylalanine during episodes of fever and infection.Also, it is difficultto use the diet to control the level phenylalanine during pregnancy andelevation can lead to a higher incidence of birth defects. Furthermore,a low phenylalanine diet itself can result in abnormalities in the fetus.

Enzyme replacement therapy

An alternative to this dietary treatment is the use of enzyme therapy toreplace the defective enzyme. However, this liver enzyme is complexand requires cofactors. It is also not stable and there is major problemrelated to large scale isolation and purification. Fortunately, there is anonhuman enzyme, phenylalanine ammonia-lyase (PAL) that can bereadily obtained fromRhodotorula glutinis (Hodgins, 1971). It does notrequire any cofactors and converts phenylalanine into trans-cinnamicacid. The liver converts the product, trans-cinnamic acid into benzoicacid that is excreted via the urine as hippurate. Trans-cinnamate hasvery low toxicity and has no adverse effects on the fetus of laboratoryanimals (Hoskins & Gray, 1982). Earlier attempts to use oral PAL didnot produce any conclusive results (Hoskins et al., 1980), most likelybecause the enzyme was destroyed by the tryptic enzymes in theintestine.


Our research on enzyme artificial cells suggests that placing PALinside artificial cells would prevent the enzyme from coming intocontact with intestinal tryptic enzymes (Fig. 6.1). At the same time,phenylalanine can enter the artificial cells to be converted by PAL.This, together with our basic research on the use of enzyme artificialcells for acatalasemia and Lesch-Nyhan disease described earlier, is anencouragement to us to look into the possible use of oral PAL artificialcells for phenylketonuria. However, one problem immediately comesto mind. The intestinal contents have amino acid concentrations thatare many times higher than those of the plasma or body fluid. How,then, are we going to extract the amino acid phenylalanine fromthe body fluid into the intestine for removal by the PAL artificialcells?

Oral PAL artificial cells for PKU based on novel theoryof enterorecirculation of amino acids

Our research has shown the presence of an extensive enterore-circulation of amino acids (Chang et al., 1989b). Pancreatic andother glandular secretions into the intestine contain large amounts ofproteins, enzymes and polypeptides. Tryptic digestion converts theseinto amino acids which are then reabsorbed back into the body as theypass down the intestine. They are converted by the liver into proteins,enzymes and polypeptides, some of which are again secreted intothe intestine. This forms a large enterorecirculation of amino acidsbetween the body and intestine. Our study shows that the dietaryprotein source of amino acids is negligible when compared withthis recirculation of amino acids. Based on this theory, it would bepossible to administer, orally, artificial cells containing one specificenzyme (e.g. phenylalanine ammonia lyase) to lower one specificamino acid (phenylalanine). PAL artificial cells given orally to PKUrats is more effective than a phenylalanine-free diet in lowering thelevel phenylalanine in the intestine, plasma and cerebrospinal fluid.Whereas PKU rats on PHE-free diet lost weight, those on oral PALartificial cells continued to grow and gained weight during the study.The following sections describe the above studies in detail.


6.10. Research Leading to Proposal ofEnterorecirculation of Amino Acids

Classical theory of the main source of amino acids inthe intestinal lumen

It is commonly thought that the tryptic digestion of proteins fromingested food constitutes the major source of amino acids in theintestinal lumen.These intestinal amino acids are then absorbed as theypass through the intestinal tract. The pancreatic and other glandularsecretions are mainly to function in supplying the required digestiveenzymes. The contributions of these secretions as a source of intestinalamino acids are not usually thought to be of major significance.

Research to test the classical theory

We give the animal nothing by mouth except sucrose and waterad libitum for 24 h. If dietary protein is indeed the major source ofintestinal amino acids, then there should be a significant decrease inthe amino acids in the intestine compared to the control group onstandard diet.

Animals on 24 h of protein-free diet

Male Sprague Dawley rats (275–330 grams) were purchased fromCharles River Co., St. Constant, Quebec. For 24 h, the control groupsreceived Purina Rat Chow (Ralston Purina, Montreal, Quebec) plus tapwater ad libitum. The protein-free diet groups received a diet of sugarcubes (Redpath Sugar Co., Montreal) and 10 g/dl glucose in tap water.Both groups of rats stayed in metabolic cages to prevent their accessto the feces.

Results of study

Figure 6.10 shows the results obtained. In the duodenum thereis surprisingly no significant difference in the amount of aminoacids between the control group on normal diet and the group onprotein-free diet. The same results are observed for the jejunum


and ileum. The concentration of amino acids decreases down theintestinal tract as amino acids are absorbed. However, there isno significant difference between the two groups throughout allsegments.There is an extremely high concentration of intestinal amino acids

as compared to the plasma concentration — more than 1000-fold inthe duodenum. Since the main source of amino acids in the intestinedoes not appear to be the proteins from ingested food, what then isthe source of the intestinal amino acids in the protein-free diet group,and indeed in the normal control group?

Proposed theory of extensive enterorecirculation ofamino acids

I thought that based on the above results, the most likely major sourceof intestinal amino acids could be the gastric, pancreatic, intestinaland other intestinal secretions (Fig. 6.9). Tryptic digestion converts thelarge amount of proteins, enzymes, polypeptides and peptides in thesecretions into amino acids. These amino acids are then reabsorbedby the body as they pass through the intestine. Continuing pancreaticand intestinal secretions into the intestine constitutes the source ofamino acids. This forms a large enterorecirculation of amino acidsbetween the body and intestine. Since the protein-free diet did not alterthe intestinal amino acid concentration, the dietary source of aminoacids was negligible when compared to that from the pancreatic andintestinal secretions.In enterorecirculation, the reabsorption of amino acids from

the intestine into the plasma could be based on standard aminoacid transport mechanisms. The transport of amino acids fromthe gastrointestinal lumen into the plasma has been extensivelyinvestigated (Stevens et al., 1984). Amino acids are generallytransported across membranes by passive or active carrier transportmechanisms (Berteloot et al., 1982). Stevens et al., (1984) added tothis classification a non-saturable diffusion PHE transport system thatcontributes significantly to the transport of phenylalanine in both thebrush border and the basolateral membranes.


Fig. 6.9. Upper: Most of amino acids in the intestine do not come fromingested food. Lower: Enzymes, proteins and peptides in pancreatic and GIsecretions, are digested by tryptic enzymes in the intestine into amino acidsthat are reabsorbed and converted by the liver into enzymes, proteins andpeptides that are again secreted into the intestine — enterorecirculation.


6.11. Oral Enzyme Artificial Cells for Genetic EnzymeDefects that Result in Elevated Systemic AminoAcid Levels

As discussed above, very surprisingly and unexpectedly, our resultshows that the major source of amino acids in the intestine appearsto be the extensive enterorecirculation of amino acids. Immediately,this raises the possibility of using this for selective depletion of specificbody amino acids. This is somewhat along the line of using bindersto remove the enterorecirculation of bile acid, thereby lowering thesystemic cholesterol level. However, unlike bile acid, there are 26amino acids in the body and usually only one amino acid needsto be selectively depleted. This is where the specificity of artificialcells microencapsulated enzyme plays a part. Oral administrationof artificial cells containing one specific enzyme (e.g. phenylalanineammonia lyase) should be able to remove only one specific aminoacid (e.g. phenylalanine) (Fig. 6.9).

General design for study of use of oral PAL artificial cells inPKU rats

In this study, we will determine whether oral PAL artificial cells candeplete a specific amino acid, phenylalanine, from PKU rats.This studywill also serve as a further test of the enterorecirculation theory. Thisstudy consists of the following groups: normal rats; PKU rats on normaldiet; PKU rats on high protein diet; PKU rats on PHE-free diet; andPKU rats on normal diet treated with oral PAL artificial cells. If there isextensive enterorecirculation, the intestinal PHE level should be aboutthe same as for the PKU rats on normal diet, PKU rats on high proteindiet and PKU rats on PHE-free diet. Furthermore, PKU rats on normaldiet treatedwith PAL artificial cells should have very lowphenylalaninein the intestine. In addition, if there is an extensive enterorecirculationof amino acids, then the treatment using oral PAL artificial cells shouldbe more effective than that using PHE-free diet alone in lowering thesystemic PHE levels (Bourget and Chang, 1985, 1986). The results tobe discussed in detail later is summarized in Fig. 6.10.


Fig. 6.10. Upper : Oral administration of enzyme artificial cells to removeone specific amino acid. Lower (left column, day 1 and right column, day 7):PHE: Phenylalanine levels; PKU phenylketonuria rats; HP high protein diet;+ diet low PHE diet: PAL AC phenylalanine ammonia lyase artificial cells.


Methods of preparation of catalase artificial cells andexperiment procedures

These are described in detail in Appendix II.

Phenylketonuria rat model

We used the rat model based on the administration of phenylalaninehydroxylase inhibitors (Anderson & Guroff, 1974). Male SpragueDawley rats, 42-day-old, (Charles River, Canada) were used to preparethe PKU rat model. The average weight of the 42-day-old rats was150 g on day 1 of the experiment. They are kept in a controlled 12 hlight/dark environment with food and water ad libitum. The rats areseparated at random into five groups: 1) Normal rats receiving a normaldiet; 2) PKU rats receiving a normal diet; 3) PKU rats receiving ahigh-protein diet; 4) PKU rats receiving a PHE-free diet (Zeigler Bros.,Inc. (USA)) and 5) PKU rats receiving a normal diet plus oral PALartificial cells — dose of 0.8ml containing 10 unit PAL/ml. Theseartificial cells are prepared by the updated method described in theAppendix under “Updated Methods for the Preparation of ArtificialCells” using collodion membrane artificial cells. Phenylalanine wasobtained from the plasma, cerebral spinal fluid and intestines justbefore PKU induction and 7 days after induction following the aboveregimes. The samples were analyzed for amino acids using highperformance liquid chromatography.

Intestinal phenylalanine

The intestinal PHE concentrations in the normal rats receiving a normaldiet did not show any significant changes in the intestine from day 1to day 7 (Fig. 6.10). In PKU-induced rats receiving a normal diet, therewere significant increases in the PHE concentrations (Fig. 6.10). Whatis most striking is that even in the PKU-induced rats receiving PHE-freediets, there were also significant increases in the PHE concentrations inthe duodenum, jejunum and ileum (Fig. 6.10). This is so even thoughno PHE was given by the oral route. This serves to further supportthe theory of enterorecirculation of amino acids from pancreatic


and other glandular secretions. PKU-induced rats receiving a normaldiet and oral PAL artificial cells showed a dramatic decrease in theintestinal PHE especially in the duodenum. This seems to show thatPAL artificial cells enzymatically degraded most of the PHE in theintestine (Fig. 6.10). This lends further support to the proposal that PALartificial cells remove PHE from the enterorecirculation.

Plasma PHE levels in PKU rats

Normal rats receiving a normal diet showed no significant changes inplasma PHE levels during the 7 days (Fig. 6.10). PKU rats receiving anormal diet or high-protein diet showed an elevated PHE level whichis more than 10-fold compared to normal rats ( Fig. 6.10). By the 7thday, the plasma levels of PKU rats receiving PHE-free diet was stillsignificantly higher than those at the normal control rats (Fig. 6.10). Onthe other hand, by the 7th day, the plasma PHE levels in PAL artificialcells-treated PKU rats were not significantly different from those of thenormal control rats (Fig. 6.10).

CSF cerebrospinal fluid in PKU rats

The normal rats showed no significant changes in the CSF PHE levelfrom day 1 to day 7 (Fig. 6.10). The PKU rats on normal diet orhigh-protein diet showed a significantly elevated level of CSF PHE(Fig. 6.10). On day 7, PKU rats receiving PHE-free diet showed asignificantly higher level of CSF PHE when compared to day 1 or thecontrol (Fig. 10). However, the level is significantly lower than that inthe PKU rats receiving a normal diet or high-protein diet. With PALartificial cells given to PKU rats on normal diet, the PHE levels inthe CSF were not significantly different from those of the normal rats(Fig. 6.10).

Growth as shown by body weight changes

Normal rats receiving a normal diet grew normally as shown bya significant increase of body weight after 7 days (Fig. 6.11). PKU


Normal PKU PKU+ HP

PKU+ diet

PKU+ PAL AC

Bo

dy

we

igh

t (g

ram

s)

200

0

7 days after PKU inductionBefore PKU induction

EFFECTS OF DIFFERENT REGIMES ON THE GROWTH OF PKU RATS

Fig. 6.11. Effects on the growth of rats followed for 7 days. From left to right:(1) normal rats; (2) PKU (phenylketonuria) rats on normal diet; (3) PKU ratson (HP) high-protein diet; (4) PKU rats on low PHE diet (diet); and (5) PKUrats receiving (PAL AC) phenylalanine ammonia lyase artificial cells.

rats receiving a normal diet or high-protein diet showed a significantdecrease in body weight. Furthermore, there was a decrease in muscletone over the same period of time. PKU rats receiving a PHE-free dietlost significantly more weight than PKU rats on normal diet. They alsoshowed a loss of muscle tone. PKU rats on a normal diet treatedwith oral PAL artificial cells had a small but significant increase inbody weight. Abnormal signs such as piloerection, aggression andhyperactivity, observed in PKU-induced rats, were no longer presentafter 5 days of treatment with PAL artificial cells.We then looked at the daily changes in body weight, comparing

normal rats, with PKU rats on PHE-free diet and PKU rats receivingdifferent amounts of PAL artificial cells.As shown in Fig. 6.12, normal rats receiving a normal diet showed

a steady and significant increase in body weight throughout the 7-dayperiod. PKU rats receiving a normal diet but no PAL artificial cells,showed a significant decrease in body weight, and in addition, adecrease in muscle tone over the same period. PKU rats on a normal


Fig. 6.12. Growth of PKU rats receiving different amounts of PAL artificialcells compared to normal rats. (modified from Bourget and Chang, 1986).

diet treated with 5 units of oral PAL artificial cells lost weight initially.However, they started to gain weight in the next 4 days at the same rateas the normal control rats. This corresponded to the time it took for theelevated plasma PHE level to return to normal. Abnormal signs such aspiloerection, aggression and hyperactivity, observed in PKU-inducedrats, were no longer present after 5 days of treatment with PAL artificialcells. PKU rats on a normal diet receiving lower doses (1.0 or 2.5 units)of PAL in artificial cells lost weight at the same rate as the PKU rats ona normal diet. Those receiving 2.5 units was only slight better off.

Congenital PKU mice model

We tried to use a congenital mouse PKU model (Shedlovsky et al.,1993) for the above studies. However, the small size of the very youngmice and the tendency for “hairy balls” to accumulate in the stomach


of these mice, has made it difficult in our hands for the long term dailygastric tube administration of 50μ PAL artificial cells. XOD artificialcells also tended to jam the very small diameter gastric tubes used fororal administration.

Summary discussion

The above results contribute to the scientific basis for the potentialtherapeutic uses of PAL artificial cells in PKU, and also serve tocompare the use of PAL artificial cells with a PHE-free diet. PAL inartificial cells acts on PHE diffusing into the artificial cells and convertthe substrate into transcinnamic acid. Transcinnamic acid is not toxicand is converted to hippuric acid by the liver and excreted in theurine. The artificial cells carry out their function as they move downthe intestinal lumen and finally leave the body in the feces. There is,therefore, no accumulation of artificial cells in the body. However, it isimportant to administer enzyme artificial cells together with a buffer toprotect them during their passage through the highly acidic stomach.At a pH of 3.5 or lower, there will be breakage of the membranes ofthe artificial cells.The safety of oral administration of enzyme artificial cells has been

shown. Thus, we have given oral artificial cells containing xanthineoxidase to lower the systemic hypoxanthine in experimental therapy inLesch-Nyhan disease (Chang, 1989a; Palmer & Chang, 1989).Artificialcells containing urease and ammonium adsorbent have also beengiven to animals (Chang, 1972a) and patients in clinical trials withno adverse effect reported (Kjellstrand, 1981).Before using this in human phenylketonuria, a key challenge has

to be solved — phenylalanine ammonia lyase is extremely expensive.There are two groups developing recombinant phenylalanine ammonialyase that is less costly (Sarkissian et al., 1999; Liu et al., 2002).

Oral enzyme artificial cells to depleteother amino acids

The proposal of oral enzyme artificial cells based on the new findingof enterorecirculation of amino acids could also be studied for


other genetic enzyme defects. These include the removal of tyrosinein tyrosinemia, histidine in histidinemia, and others. In addition,the basic results obtained here can be useful for analyzing thefeasibility for use in other conditions. One example is the use oforal administration of artificial cells containing asparaginase to lowerasparagine in leukemia. Asparaginase is used as an adjunct treatmentin chemotherapy, but it has side effects when injected intravenouslyon a repeated basis. Oral use of asparaginase artificial cells if effective,would avoid the adverse effects related to intravenous injection.Removal of glutamine is another example. The use of oral tyrosinaseartificial cells can effectively lower the systemic tyrosinase level.As described in the next chapter, we are studying this for potentialapplication in melanoma, a skin cancer that depends on tyrosine forgrowth.

Documents

Use of EnzymeArtiﬁcial Cells for Genetic Enzyme … Enzyme Defects that Increase Systemic Substrates toToxic Levels ... enzyme may result in immunological and ... enzyme catalase